1. Berkeley Segmentation Dataset
   • The Segmentation Dataset - Martin 03
   • Segmentation Performance Evaluation- Martin 03
2. Normalized Cuts Algorithms/Theory/Analysis
   • Normalized Cuts justified from discrete 2-way criterion - Shi and Malik 00
   • Perturbation theory for a k-way algorithm - Ng,Weiss and Jordan 02
   • Multiclass justified from discrete k-way critereon - Yu and Shi 03
   • Gradient ascent for learning affinity functions - Bach and Jordan 04
3. Bottom-up Segmentation Cues
   • Learning optimal boundary detection - Martin, Fowlkes and Malik 04
   • Learning optimal affinities - Fowlkes, Martin and Malik 03
4. Incorporating Top-Down Recogntion Cues
   • Optimizing nomralized cuts with grouping constraints - Yu and Shi 04
   • Incorporating top-down recognition cues - Yu and Shi 03
5. Multi-scale Computation
   • Multiscale optimizaition of normalized cut - Sharon, Brandt and Basri 00
   • Multiscale intensity cues - Sharon, Brandt and Basri 01
   • Aggregating multiscale texture cues - Galun, Sharon, Brandt, and Basri 03
6. Computational Effeciency
   • Nystrom approximation - Fowlkes, Belongie, Chung and Malik 04